EPIA'2011
ISBN: 978-989-95618-4-7
Thus. the purpose of tliis study is to identify groups of similar gamę moves. By identifying these move groups. we can characterize an opponent by knowing the group of moves that it uses most.
The goals of this work are:
■ Process a significant amount of poker gamę logs between human players, from casino clients:
■ Extract information about the events of the games presented on the gamę logs:
■ Use clustering algorithms to identify groups of common gamę moves;
■ Measure the players' relative freąuency of use of the identified move groups in order to identify common strategies.
The rest of the paper is organized as follows. Section 2 describes the gamę of Poker and the variant of Poker that was studied - No Limit Texas Hołd em. Section 3 describes related work about some approaches followed earlier to create Poker agents as well as the description of the type of clustering algorithms that wcre used in this work. Section 4 presents a generał overview of tliis work. including the steps of development from the extraction of Poker liands to the definition of strategies. Section 5 presents the steps that were followed to extract information from Poker games. present on the gamę logs. Section 6 presents the usage of clustering algorithms to dcterminc types of actions in Poker games. Section 7 presents the usage of clustering algorithms to determine types of strategies. based on the frequency of each type of action. Finally, section 8 presents the paper main conclusions and some pointers for futurę work.
Poker is a generic name for literally hundreds of games. but they all fali within a few interrelated types [3|. It is a card gamę in which players bet tliat their liand is stronger than the hands of their opponents. All bets go into the pot and at the end of the gamę the player with the best hand wins. Another way of winning is making the other players forfeit the hand and therefore being the last standing player.
2.1 No Limit Tcxas Hold’em
No Limit Texas Hołd em is a Poker variant that uses community cards. At the beginning of eveiy gamę, two cards are dealt to each player. A dealer player is assigned and marked with a dealer button. The dealer position rotates eloekwise from gamę to gamę. After that. the two players to the left of dealer post the blind bets (small-blind and big-blind). The player that starts the gamę is the one on the left of the big blind. One e.xample of an initial table configuration is on the Figurę 1. The dealer is the player F and the smali and big blind players are respectably A and B seats.
71